Machine learning may better detect autism when neuroimaging data include women
A slew of studies have sought to use machine learning to detect autism from brain images. The source of that data matters; including women and using multiple imaging technologies may increase the sensitivity of certain machine-learning models. Hover over a dot to see additional information.
A forest plot showing the pooled sensitivity of various machine learning models, when trained on different types of data. Neuroimaging data that includes both males and females, for example, is more sensitive at spotting signs of autism than data that includes males only.
